Pittsburgh's medical community creates a lot of data, and embedded within that data are potential answers to important questions. The Pittsburgh Health Data Alliance -- a collaboration of UPMC, the University of Pittsburgh and Carnegie Mellon University -- sifts through that data. This month, they announced a new sponsor in that effort: Amazon's computing subsidiary.
"Amazon has great cloud infrastructure resources and tools that are available for doing research using techniques like machine learning," said UPMC Enterprise's Zariel Johnson, who manages daily activities of the Alliance. "These are highly useful to the researchers that we're working with as part of the [alliance]."
The alliance uses machine learning for some of its projects to sift through huge amounts of data from biomedical research and the health care field. Researchers are using the computing method to answer questions such as what factors impact readmission rates after surgeries, or how a person's genetic makeup influences influences their likelihood of developing a certain type of cancer.
"Computers are extremely useful when we're talking about health care and biomedical research due to the amount of data that's being produced," Johnson said. "There are simply too many variables for a human to ever be able to sift through in multiple lifetimes."
Criticism of "big data" has persisted for years, with some coming out against the collection of personal information by large digital companies, including Amazon. Johnson said the company's sponsorship of some PHDA projects will not grant it access to any data involved, and all the data follows national privacy standards.
"We take every step along the way possible in order to safeguard the data," she said.